From prompts to products: One year of Responses | OpenAI Developers

This post details how developers are using the Responses API to build and operate AI agents, focusing on two key areas: 1) Raindrop AI's system for detecting and fixing failures in AI agents in production, which leverages the Responses API for background analysis, tool sharing, and surfacing unusual behavior for debugging. 2) Repo Prompt's system for deep reasoning workflows on complex data, which separates context gathering from deep reasoning using agent workflows and the Responses API for orchestration, background jobs, and observability. It also highlights Collxn's conversational interface for vinyl record collectors, using the Responses API for tool calling to interact with Discogs data.